Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does so well: it flags paid cost, real-world irreversible side effect on a real person, the consent requirement, the absence of a sandbox, and exact idempotency-key retry semantics. It discloses not just what happens but the consequences for the agent's decision-making.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The critical warning ('HAS A REAL-WORLD SIDE EFFECT') is front-loaded, and every following sentence — consent, sandbox, idempotency — adds a distinct fact an agent needs. No filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the absence of annotations and output schema plus a nested body requiring get_schema, the description covers the safety and retry behavior thoroughly. The one minor gap is that it doesn't say what comes back (e.g., a receipt), which would help an agent chain to get_receipt.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the schema already documents slug, body, and idempotency_key. The description repeats the idempotency-key semantics almost verbatim from the schema and reiterates the consent block, adding little new meaning beyond the baseline.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

It states the concrete effect — 'it puts a message in front of an actual person' — which is more specific than the generic name 'call_action'. However, it never names the operation as 'sends a message via the selected action slug', and it doesn't differentiate itself from siblings like sandbox_call or get_receipt by name.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

'There is no sandbox for this — a simulated send cannot be told apart from a real one' is a clear usage exclusion and implicitly steers the agent away from sandbox_call, while the idempotency note tells it how to handle retries. The alternative tool isn't explicitly named, so it's clear context rather than full when/when-not/alternatives guidance.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.